Article Text

Download PDFPDF

Are there risk factors in alpine skiing? A controlled multicentre survey of 1278 skiers
  1. R M Hasler1,
  2. S Dubler1,
  3. L M Benneker2,
  4. S Berov3,
  5. J Spycher3,
  6. D Heim4,
  7. H Zimmermann1,
  8. A K Exadaktylos1
  1. 1
    University of Bern, Department of Emergency Medicine, Inselspital, Bern, Switzerland
  2. 2
    University of Bern, Department of Orthopedic Surgery, Inselspital, Bern, Switzerland
  3. 3
    Spital Interlaken, Department of Orthopedic Surgery, Interlaken, Switzerland
  4. 4
    Spital Frutigen, Department of Surgery, Frutigen, Switzerland
  1. Correspondence to Dr L Benneker, Department of Orthopedic Surgery, Inselspital Bern, University Hospital, CH – 3010 Bern, Switzerland; lorin.benneker{at}insel.ch

Abstract

Objective: To analyse risk factors in alpine skiing.

Design: A controlled multicentre survey of injured and non-injured alpine skiers.

Setting: One tertiary and two secondary trauma centres in Bern, Switzerland.

Patients and methods: All injured skiers admitted from November 2007 to April 2008 were analysed using a completed questionnaire incorporating 15 parameters. The same questionnaire was distributed to non-injured controls. Multiple logistic regression was performed. Patterns of combined risk factors were calculated by inference trees. A total of 782 patients and 496 controls were interviewed.

Results: Parameters that were significant for the patients were: high readiness for risk (p = 0.0365, OR 1.84, 95% CI 1.04 to 3.27); low readiness for speed (p = 0.0008, OR 0.29, 95% CI 0.14 to 0.60); no aggressive behaviour on slopes (p<0.0001, OR 0.19, 95% CI 0.09 to 0.37); new skiing equipment (p = 0.0228, OR 59, 95% CI 0.37 to 0.93); warm-up performed (p = 0.0015, OR 1.79, 95% CI 1.25 to 2.57); old snow compared with fresh snow (p = 0.0155, OR 0.31, 95% CI 0.12 to 0.80); old snow compared with artificial snow (p = 0.0037, OR 0.21, 95% CI 0.07 to 0.60); powder snow compared with slushy snow (p = 0.0035, OR 0.25, 95% CI 0.10 to 0.63); drug consumption (p = 0.0044, OR 5.92, 95% CI 1.74 to 20.11); and alcohol abstinence (p<0.0001, OR 0.14, 95% CI 0.05 to 0.34). Three groups at risk were detected: (1) warm-up 3–12 min, visual analogue scale (VAS)speed >4 and bad weather/visibility; (2) VASspeed 4–7, icy slopes and not wearing a helmet; (3) warm-up >12 min and new skiing equipment.

Conclusions: Low speed, high readiness for risk, new skiing equipment, old and powder snow, and drug consumption are significant risk factors when skiing. Future work should aim to identify more precisely specific groups at risk and develop recommendations—for example, a snow weather index at valley stations.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

“Citius, altius, fortius”—faster, higher, stronger. This is not only an ancient Olympic slogan, but a trend in modern winter sports as well. Skiers now progress to more advanced skills and faster speeds than ever before, mainly because of better prepared slopes and high-tech equipment.1 This has allowed even moderately trained sportsmen and sportswomen to race the slopes like professionals. High-energy falls and collisions may result when skiers underestimate the risks they take. The consequence may be severe injuries, such as head and spine trauma, which can be associated with global functional impairment and give rise to long-term disability.2

Although there are many reports in the literature describing the epidemiology and biomechanics of ski injuries, data on risk factors are very limited. Previous studies have looked at a few single risk factors or risk factors related to specific injuries.3 4 5 6 7 8 9 The large number of injuries related to snow sports, in combination with the rising costs of healthcare, prompted our institution to look for ways to cooperate with other trauma centres and major insurance companies, with the aim to become more efficient at preventing injuries by detecting general risk factors in alpine skiing.

Several reports have documented that the frequency and severity of injuries are linked to experience and speed.10 11 However, our impression from years of experience and from the results of a pilot study is that many skiers suffer injuries at slow or medium speeds and that the probability of injury to alpine skiers may be linked to a concatenation of several risk factors, such as age, gender, experience, use of protectors or weather conditions (unpublished data from an internal survey).

To establish whether our impressions could be substantiated, we conducted a prospective controlled multicentre survey on 1278 injured and non-injured skiers during the season 2007/2008, with the aim of defining statistically significant risk factors and risk factor combinations in alpine skiing.

Patients and methods

Setting and interventions

Injured patients were interviewed at the emergency department, Inselspital, University Hospital Bern, a level I trauma centre (the only tertiary trauma unit, with a catchment area of 1.8 million people, serving the alpine region of central Switzerland, the Bernese Oberland) and at the emergency departments of Interlaken and Frutigen Regional Hospitals, both level II trauma centres. The control group consisted of subjects interviewed at valley stations of six different ski resorts belonging to the catchment area of the three hospitals. Between 1 November 2007 and 15 April 2008, a questionnaire, incorporating 15 parameters of potential risk factors in alpine skiing, was distributed to injured and non-injured skiers.

Patients and controls

The patient group contained 782 injured alpine skiers of any age, independent of skiing level and experience, who were admitted to the trauma centre of one of the three hospitals. The control group contained 496 non-injured alpine skiers of any age, independent of skiing level and experience. Participants were interviewed using the standardised questionnaire at the emergency departments (patients) or when coming off the slopes (controls). Questionnaires were available in multiple languages. If younger patients (<12 years) were not able to answer the questions, accompanying parents or relatives were interviewed on their behalf. Participation in the study was voluntary and anonymous; confidentiality was granted.

Data were collected, stored, analysed and shared according to the ethics committee standards of the Inselspital Bern.

Study measures

We defined 15 primary outcome measures as possible risk factors: patient/control characteristics (age, gender, years of experience in alpine skiing), behavioural aspects (readiness for risk and speed, abstinence from alcohol or drugs while skiing, duration of warm-up), equipment (use of protectors, age of skis, seasonal checking of skis) and external conditions (slope and snow conditions, weather and visibility, experience of aggressive behaviour of other skiers or snowboarders). Readiness for risk and speed were measured on a self-reported visual analogue scale (VASrisk, VASspeed) from 1 to 10 (1 implying minimal speed or risk and 10 maximum speed or risk). We included three different types of protectors: (a) helmets, (b) back protectors and (c) wrist protectors.

Seasonal checking of skis was defined as a control by a sports store specialist in the current winter season. We analysed three types of slope conditions as assessed by the individual skiers—(a) hard and icy, (b) soft and powdery or (c) slushy—as well as three types of snow conditions, also individually assessed: (a) fresh snow, (b) old snow or (c) artificial snow. Weather and visibility while skiing were classified as follows: (a) sunny weather and good visibility or (b) cloudy weather and bad visibility. The parameters of aggressive behaviour on slopes included individually reported experiences of (a) being deliberately kicked by another person, (b) being pressurised by another person or (c) having witnessed an argument. Furthermore, we measured combinations of two or more parameters. Therefore, a model was set up using a conditional inference tree, in order to describe typical combinations of predictors (appendix I).

Statistical analysis

To detect the influence of the risk factors defined above, single and multiple logistic regression were performed, and odds ratios (ORs) with corresponding 95% CI were reported for all of these parameters. Because multiple logistic regression analysis based on a large study population was used for this survey, patient–control matching was not necessary.

For ordinal or metric variables, ORs were expressed as the ratio of the odds from the 3rd to the 1st quartile of the corresponding distribution. p<0.05 was considered significant.

All evaluations were calculated with R version 2.7.0 (a language and environment for statistical computing; appendix II). Details of conditional inference trees are given in appendix III.

Results

A total of 782 patients and 496 controls were identified between 1 November 2007 and 15 April 2008 who satisfied the inclusion criteria. Nineteen alpine skier controls were excluded because of incomplete data.

Controls had a mean age of 35 years; 52% were female and 48% were male. Patients had a mean age of 40 years; 43% were female and 57% were male.

The mean experience of alpine skiing was 22 years in the control group and 20 years in the patient group.

Single and multiple logistic regression analysis were performed (figs 1–3). Multiple logistic regression analysis showed a significant correlation with the patient group for high risk and low speed (table 1). There was a significant relation between no experience of aggressive behaviour of others on the slopes and injury. The use of new skiing equipment and performance of warm-up actions were significantly more common in the patient group. Old snow rather than artificial or fresh snow and powder snow rather than slushy snow were more often found in the patient group. Consumption of alcohol and abstinence from drugs while skiing was more common in the group of injured skiers. A trend (within the 90% CI) was detected for bad weather/poor visibility and seasonal checking of skiing equipment (table 2). There were no significant findings for the parameters age and gender, years of experience of alpine skiing, use of wrist and spine protectors, and hard versus powder snow (table 3).

Figure 1

Single categorical parameters for the patient and control group.

Figure 2

Box plot diagram of the single metric parameter “speed” showing higher readiness for speed for the control group.

Figure 3

Odds ratios of multiple logistic regression analysis for each parameter. The white part of the bar indicates 95% CI or more.

Table 1

Parameters significant for the patient group

Table 2

Trend (within 90% CI) for the patient group

Table 3

Parameters not significant for the patient group

Conditional inference trees

Application of the inference trees of combined risk factors revealed three “typical” patient groups for alpine skiing:

1. Warm-up of 3–12 min, VASspeed >4 and bad visibility

2. VASspeed between 4 and 7, icy slopes and not wearing a helmet

3. Warm-up >12 min and new equipment

Figures 4 and 5 illustrate the confidence inference tree. Of these variables, warm-up had the greatest influence on variability, followed by speed.

Figure 4

Inference tree 1 of risk factor combinations in alpine skiers. Two examples of conditional inference trees used to postulate groups at risk by a combination of the most significant parameters. Information for interpretation of conditional inference trees can be found in appendix IV.

Figure 5

Inference tree 2 of risk factor combinations in alpine skiers. Two examples of conditional inference trees used to postulate groups at risk by a combination of the most significant parameters. Information for interpretation of conditional inference trees can be found in appendix IV.

Discussion

There has been a lot of speculation that the incidence of serious injuries in snow sports may have changed over the past 10 years.12 13

In this large study, we identified several single parameters that were concentrated in the patient group (fig 1). We must, nevertheless, be cautious when declaring these as risk factors, as they are significant only when looked at as a single parameter without respect to the distribution of all the other parameters in the patient and control group. Therefore we applied multiple regression analysis (fig 3). This method differs from the conventional approach of a matched patient–control design, which is mainly used in smaller study populations. However, selection bias is not completely excluded. What is more, a combination of multiple risk factors may lead to injuries. To overcome this problem, we applied typical inference trees and successfully identified three groups of injured skiers. This has the character of hypothesis generation, as the three groups were not postulated in advance (figs 4 and 5).

Groups of injured skiers and warm-up

The first group represents the ambitious, athletic skier who performs a long warm-up and has new equipment which allows aggressive and fast skiing, where small mistakes can have fatal consequences.

The second group may include skilled and ambitious sportsmen who spend 3–12 min on warming-up. This large cohort is especially at risk at high speeds and poor visibility. These two groups are responsible for the result that a warm-up of >3 min before skiing is over-represented in the patient group, with an OR of 1.8. Of course, this result does not imply that the warm-up should be omitted, as this study did not link the performance of warm-up with specific types of injury. Therefore the over-representation of warming-up has to be interpreted because of the limitations of the study design.

The third group possibly represents the less experienced and more careless skier. Injured skiers from this group often wear no protective gear and are at high risk at moderate speeds and in critical slope conditions (the highest risk is for the combination icy slopes and not wearing a helmet).

Equipment and readiness for risk and speed

Another surprise was the finding that skiers with new equipment have a higher risk of being injured. As we have stated in the introduction, this might be explained by a mismatch between the abilities of the skier and his/her equipment. It is important for skiers to be aware that simply checking the equipment regularly does not mean that skiing is safe. Rental stations and ski shops should display such information.

Skiing equipment has developed to more progressive side cuts, and new materials and technologies are highly resistant to torsion forces, allowing optimal energy transfer to the edges of the ski. Thereby higher speeds and a more risky carving style are possible, even for skiers of limited skill and training.14 Increased speed and larger turns tend to restrict space on the slopes, as each individual skier demands more.

Snow and slope conditions

Ski resorts have invested large sums in the production of artificial snow, and this often leads to hard and then icy slopes.

As expected, snow conditions do play a significant role. Skiers should be aware of this. Snow and slope conditions should be graded, and skiers should be informed at the valley stations. A snow quality/weather/injury risk grading would be ideal.

Alcohol and drug consumption

Traditionally, alcohol and drug consumption have been regarded as risk factors for skiing injuries.15 16 Alcohol consumption, the so-called “après ski”, has become part of ski and snowboard culture, regardless of age and gender. This is, however, not consistent with the impression we gained. The over-representation of alcohol consumption in the control group (OR −1.99) was most surprising. Either alcohol consumption is more moderate than previously assumed or the questions were not answered honestly.17 Another possible explanation is that our study was performed at the same time as a new, more restrictive drink and driving limit was implemented in Switzerland, with lowering of the legal blood alcohol limit. This has resulted in some positive effects on the drinking habits of a large proportion of the Swiss population. A similar effect on skiing accidents might be possible, as up to 70% of skiers travel by car.

This is in contrast with drug consumption, which was shown to be a significant risk factor (OR 1.78) for injuries. Consumption of soft drugs (mainly cannabis in Switzerland) does not attract heavy sanctions by law and is, in general, well tolerated by society and very popular among youngsters.

Aggressive behaviour on slopes

Many skiers felt intimidated by others, but we could not detect any relation between unsocial and aggressive behaviour and injuries. This indicates that aggressive behaviour alone is not a major factor in causing ski accidents.

Conclusions

We conclude that several risk factors and combinations of risk factors exist, but, when drawing conclusions, one has also to be aware of differences in risk profiles. Future work should be aimed at confirming the existence of the three groups at risk and developing recommendations or interventions specific to each group. Snow and slope conditions should be graded, and skiers should be informed at the valley stations. A snow quality/weather/injury risk grading would be ideal. As for road traffic, where only a few crashes are caused by drunk and unsocial drivers, most skiing accidents can be attributed to interactions between several factors. Aware and careful skiing is probably the best way to prevent injuries.

Acknowledgments

We acknowledge Mrs Kathrin Dopke MPH, study coordinator, Department of Emergency Medicine, Inselspital, University Hospital Bern, Switzerland, and Rodney Yeates PhD, for editing the English.

Appendix I

15 primary outcome measures:

Patient and control characteristics

(1) Age: Years

(2) Gender: F/M

(3) Experience in alpine skiing: Years

Behavioural aspects

(4) Readiness for risk: VAS 1–10 (1 implying minimal speed or risk and 10 a maximum of speed or risk)

(5) Readiness for speed: VAS 1–10 (1 implying minimal speed or risk and 10 a maximum of speed or risk)

(6) Abstinence from alcohol while skiing: YES/NO

(7) Abstinence from drugs while skiing: YES/NO

(8) Duration of warm-up before the first run: Minutes

Equipment

(9) Use of protectors:

(a) helmet

(b) spine protector

(c) wrist protector

(10) Age of skis: Years

(11) Seasonal checking of skis by a sports store specialist in the current winter season: YES/NO

External conditions

(12) Slope conditions:

(a) hard/icy

(b) soft/powdry

(c) slushy

(13) Snow conditions:

(a) fresh snow

(b) old snow

(c) artificial snow

(14) Weather and visibility:

(a) sunny weather and good visibility

(b) cloudy weather and bad visibility

(15) Experience of aggressive behaviour of other skiers:

(a) being deliberately kicked by another person

(b) being pressurised by another person

(c) having witnessed an argument

Appendix II

R Development Core Team (2008). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org.

Appendix III

Torsten Hothorn, Kurt Hornik and Achim Zeileis (2006). Unbiased recursive partitioning: a conditional inference framework. Journal of Computational and Graphical Statistics 2006;15(3):651–74.

Appendix IV

1. Berk RA. Statistical learning from a regression perspective series: Springer series in statistics. 2008, XVIII, 360 pp, hardcover. ISBN: 978-0-387-77500-5. Chapter 3 Classification and Regression Trees (CART), p 103.

2. http://en.wikipedia.org/wiki/Decision_tree_learning.

3. Original research on inference trees: http://statmath.wu.ac.at/∼zeileis/papers/Hothorn+Hornik+Zeileis-2006.pdf: Torsten Hothorn, Kurt Hornik and Achim Zeileis (2006). Unbiased recursive partitioning: a conditional inference framework. Journal of Computational and Graphical Statistics 2006;15(3):651–74.

What is already known on this topic

  • Although there are many reports in the literature describing the epidemiology and biomechanics of ski injuries, data on risk factors are very limited.

  • Previous studies used a conventional approach of a matched patient–control design and looked at only a few single risk factors or risk factors related to specific injuries.

What this study adds

  • Multiple regression and conditional inference tree analysis has not been used before in such a study.

  • Not only one but several risk factors and their combinations were identified: low speed, high readiness for risk, new skiing equipment, old and powder snow as well as drug consumption.

REFERENCES

View Abstract

Footnotes

  • Competing interests None.

  • Provenance and Peer review Not commissioned; externally peer reviewed.